Tata Communications — a global network learning to sell software, priced as if it already has
Tata Communications Limited
Tata Communications — a global network learning to sell software, priced as if it already has
The Pulse
Tata Communications owns some of the least glamorous and most irreplaceable plumbing on the planet — the only wholly-owned fibre-optic ring that circles the globe under the sea, and the world’s largest wholesale voice network — and is in the middle of a long, deliberate attempt to stop being judged as plumbing. The pitch is a “digital fabric”: connectivity plus cloud, security, IoT and customer-interaction software, sold as a bundle to the world’s biggest companies. Revenue has crept back to ₹24,803 crore in FY26 (+7.3%), digital products now make up more than half of data revenue for the first time, and the company throws off real cash. But the transformation costs money before it pays: operating margins have slid from 25% to 19%, return on capital from 24% to about 15%, and the digital portfolio everyone is excited about actually lost roughly ₹900 crore in FY25. A brand-new CEO arrived in May 2026 and promptly declined to reaffirm his predecessor’s growth targets. The stock, meanwhile, trades at 50 times earnings and 16 times book — priced for the software company it is trying to become, not the carrier it still mostly is.
The Business
Strip away the jargon and Tata Communications is a toll operator on global data and voice. It built (or bought) the physical infrastructure — subsea cables, network points-of-presence in 200-odd countries, data-centre interconnects — and then rents capacity and services on top of it to roughly 300 of the Fortune 500. About 60% of revenue comes from outside India. The legacy half of this is two melting ice cubes: wholesale voice (carrying international calls, a business in permanent structural decline) and “core connectivity” (selling bandwidth, where prices erode and 10–15% of revenue churns away each year). The growth half — the “digital portfolio” — is everything richer: cloud networking, security, the Kaleyra-derived messaging-and-interaction business, and an AI cloud.
What genuinely makes the company distinctive sits in two places. First, the network itself: that subsea ring and >40% share of India’s data-centre-to-data-centre connectivity are hard to replicate, and the relationships are sticky — eight of India’s top ten banks run on its network fabric, and customers using three or more “fabrics” rarely leave because ripping out live infrastructure is the corporate equivalent of changing a plane’s engine mid-flight. Second, a quiet jewel called the Campaign Registry — an anti-spam, messaging-compliance data business acquired with Kaleyra that earns roughly 50% segment margins on almost no capital. The rest of the digital portfolio is still proving it can make money. The promoter is the Tata group, steady at 58.86% for years; the government, which once owned the whole thing as VSNL, exited entirely in FY21. Worth noting in the ownership tape: foreign institutions have been selling (18% down to 14.4%) and domestic ones buying.
How Management Thinks
The previous CEO, A.S. Lakshminarayanan, and his hard-nosed CFO Kabir Shakir ran the most candid telecom investor presentation you’ll read. At the June 2025 Investor Day they did something companies rarely do: quantified their own failure. They disclosed for the first time that the digital business was losing ₹900 crore, called it a ~₹2,300 crore EBITDA opportunity if fixed — and then refused to say when. (“Don’t ask me what the period is. The moment we quantify it, neck is on the block.”) They explained that buying the STT data-centre stake alone costs about 220 basis points of return on capital, and said plainly they would accept a lower near-term number for a better long-term one. The capital-allocation philosophy is genuinely disciplined for the stage: maintenance capex hard-capped at ~2% of revenue, customer-linked capex gated on the profitability of each deal, big strategic bets gated on IRR. They termed out debt at fixed rates, monetised surplus real estate, exited sub-scale businesses, and kept net debt below 2x EBITDA. Dividends are paid (about half of profit); there are no buybacks despite a history of margin-dilutive acquisitions and a falling return profile.
The honesty cuts both ways, though, and the scoreboard is mixed. Return on capital has fallen the entire time this management has run the pivot. The acquisitions that built the digital fabric (Kaleyra, Switch) diluted margins and have yet to fully “deliver back.” And reported profit is genuinely noisy — FY25’s looked great (₹1,837 crore) almost entirely because of a one-off ₹1,033 crore of other income, then FY26’s “halved” to ₹997 crore largely because that one-off didn’t repeat. Underlying segment profit was flat both years. A patient reader has to look past the headline net profit to the cash flow (₹4,479 crore from operations in FY26) and the segment tables to see what’s actually happening, which is: heavy reinvestment, not yet much payoff.
The new CEO, Ganesh Lakshminarayanan, took over in May 2026 and on his first call — 45 days in — politely refused to inherit anyone’s promises. Asked directly whether the old “double the data business” target still held, he said it was “too early to comment” and parked everything to an Investor Day after his 100 days. He reframed the goal from top-line ambition to “profitable growth” and growing absolute EBITDA, openly absorbing the bear case that EBITDA growth has been “flat to low-single-digit.” His self-diagnosis is telling: the problem, he thinks, is storytelling, not capability — “customers want us to do more; we need to do a better job explaining the full portfolio.” Whether a narrative fix solves a returns problem is the open question hanging over the stock.
Where It’s Going
The strategy is unchanged in substance even if the targets are now under review. The old aspiration was ₹28,000 crore of data revenue (a near-doubling, originally framed around FY28 but already conceded to have “moved a few quarters”), with digital rising to 65% of data revenue and margins recovering to 23–25% and ROCE back above 25% once the digital business turns profitable. The demand backdrop genuinely supports the direction: India’s data-centre capacity is expected to double, AI workloads need exactly the low-latency, data-centre-to-data-centre connectivity Tata leads in (management sizes the India DC-interconnect opportunity at $1bn-plus by 2030), and the country’s army of global capability centres is set to roughly double. Tata’s chosen position is deliberately picks-and-shovels: it is not building large language models or competing to be a hyperscaler; it sells the secure, sub-3-millisecond network around the AI — the neutral fabric between clouds, enterprises and data centres.
The tensions are equally real. The headline 9.4% revenue growth in the latest quarter was only 3.7% once you strip out a friendly dollar — management volunteered this, to their credit. A persistent and unexplained gap between strong double-digit order booking and single-digit revenue conversion has been raised by analysts for several quarters and never satisfactorily answered. Red Sea cable cuts dented the high-margin connectivity business. A fresh West Asia flare-up is postponing the live-event broadcasts (F1, MotoGP) that Tata’s media business serves. And the whole thesis still rests on the digital portfolio swinging from a ₹900 crore loss to double-digit margins — promised, plausible, but not yet delivered. A separate swing factor is the 26% stake in STT’s Indian data-centre business, which management plans to monetise via an IPO “for the right value,” with proceeds-use left open. The direction of travel is coherent; the question is purely one of execution and pace, and the new CEO has bought himself time to answer it.
The Four Checks
1. Quality & moat (gate) — 6/10. There is a real moat, but it is uneven and partly eroding. The subsea fibre ring, >40% India data-centre-interconnect share, entrenched multi-fabric relationships with the world’s largest enterprises (eight of ten top Indian banks), and the ~50%-margin Campaign Registry niche are genuinely hard to replicate. But the largest chunk of revenue — wholesale voice and core connectivity — is commoditising, with price erosion and double-digit churn, and the falling return on capital is the moat showing strain. A strong-but-contestable business, not an unassailable one.
2. Returns on incremental capital & runway — 5/10. The runway is large and open (data centres, AI connectivity, GCCs, security). The returns are the problem. ROCE is ~15% and has fallen from 24% three years ago. Capital is going almost entirely into Data Services (₹3,300 crore of ₹3,349 crore in FY26), yet that segment’s profit was flat — the heavy build is not currently earning its keep. The bull case requires digital margins to turn, lifting incremental returns; the evidence so far shows the cost of the bet, not the payoff. Moderate returns, big runway, unproven incremental economics.
3. Capital allocation for the stage — 6/10. Process is genuinely good: IRR-gated capex, fixed-rate debt, real-estate and non-core monetisation, leverage held below 2x, candid disclosure, ~50% dividend payout, and a BCG-cited #1 five-year total-shareholder-return rank among global telcos. The quibbles are not trivial: the margin-dilutive M&A, the absence of buybacks despite a depressed return profile, and a returns scoreboard that has moved the wrong way for years. Rational and disciplined for an investment phase, but the outcomes lag the process.
4. Price — 3/10. Demanding. 50x earnings (on lumpy, one-off-flattered profit) and 16x book for a business that grew sales 7.7% over five years with falling margins and returns. The more forgiving lens — EV/EBITDA of roughly 13–14x given the depreciation-heavy structure — is full rather than absurd, and the market is clearly paying for the digital-fabric narrative to come good. But the price leaves little room for the pivot to disappoint, and the new CEO has just declined to underwrite the targets that justify it.
Engine score: 17/30 (moat 6, reinvestment 5, allocation 6) — a quality infrastructure asset with a coherent strategy and disciplined, candid management, held back by an unproven reinvestment payoff and a commoditising core. Price (3) sits well below the business quality, the classic tension of a good company at a demanding valuation.
Sources
- Concalls read: Investor & Analyst Day (10 Jun 2025); Q2 FY26 results call (15 Oct 2025); Q4 FY26 results call (22 Apr 2026). Q1 FY26 and Q3 FY26 (Jan 2026) transcripts were not available — the Oct-2025 quarter was PPT-only on screener, and the Jan-2026 filing was a BSE cover letter pointing to the company site, with no transcript body.
- Annual reports read: FY24, FY25, FY26 (trimmed high-signal sections). Note: in all three AR extracts the Chairman’s and MD’s narrative prose survived only as headings; qualitative framing was reconstructed from the strategy/risk sections and the audited Note 39 segment tables, which were fully captured.
- Screener snapshot fetched: 2026-06-18 (logged-out / public; consolidated). Caveats carried through: reported net profit is materially distorted by Other Income swings (₹1,033 cr FY25; ₹926 cr in the Mar-2025 quarter), so ROE of 34% / 48% flatters a falling operating return; screener flags possible interest-cost capitalisation.
- Research subfolder (not published):
vault/Sources/Earnings/Tata Communications Ltd/